AJST — A Joint Storage & Toolkit for Transient Astronomy

AJST is a self-hostable web database and analysis toolkit for GRB afterglows and other astrophysical transients. It combines a multi-band light-curve catalog (currently 2,794 events, 257,330 photometry points, 963 redshifts) with Galactic-extinction machinery, afterglow and host-galaxy SED fitting, prompt-emission scaling relations, and everyday research tools (figure digitizer, GCN reader, synthetic spectrophotometry). The stack is a Flask + PostgreSQL backend and a build-free vanilla-JS frontend with all libraries bundled locally — clone, pip install, point at PostgreSQL, run.

• Transient registry — coordinates (decimal/sexagesimal), T0 with reference and offset metadata, redshift (spec-z / photo-z / upper limit), trigger instrument, tags, aliases, free-form JSONB extension fields.
• Unified photometry store — per-point original value and unit (mag/mJy/μJy/Jy/cgs), magnitude system (AB/Vega), upper-limit and host-subtraction flags, telescope/instrument, and reference; times normalized to Δt since T0 with MJD as the authoritative observation time.
• Editing — two-tier permissions (admin / regular users), inline table editing, batch operations, non-destructive "discard" flagging, CSV upload with column mapping, per-event literature lists with BibTeX.
• Reversible file↔database ETL (etl.py --sync / --dump / --prune) against the versioned AJST-Data layout (one JSON + one CSV per event).

Detail page — basic parameters, literature with one-click BibTeX, Aladin Lite finder chart, external-catalog parameters (T90, Epeak, Eiso, … from GBM/BAT/XRT and literature samples), and derived rest-frame quantities:

Light-curve plotting — multi-band plots with error bars, linear/log axes, flux-density or absolute-magnitude Y axes, observer/rest-frame switching, and overlay fits of empirical models (power law, broken/smoothly-broken power law, FRED pulse) with 1σ credible bands from emcee posteriors:

• Galactic extinction — per-point or whole-catalog correction with the CSFD dust map (dustmaps) and the Pei (1992) P92 curve (R_V = 3.1, dust_extinction); Vega→AB conversion per filter; results cached and re-derived automatically when coordinates or filters change.
• Afterglow fitting — MCMC fits via VegasAfterglow with a preset engine covering 216 jet-structure × circumburst-medium × host-extinction combinations plus joint constraints (FS+RS pairs, two-component jets); corner plots and chains archived per job.
• Host galaxies — per-event host coordinates, redshift, and photometry (AB/Vega/ST), with a SED-fitting tab running CIGALE/pcigale (Boquien et al. 2019) or prospector (Johnson et al. 2021); fixed-z or photo-z runs with one-click write-back of M*, SFR, age, A_V.
• Spectra — upload, overplot, offset, and download spectra; rest-frame sub-axis; TNS-style line marking (30 groups with per-group z and expansion velocity); optional extinction-corrected secondary products (CSFD + P92).
• Spectra × filters workbench — synthetic photometry of library or user-uploaded spectra through registered filter curves, with anchoring and error budgeting.
• Statistical relations — Amati, Yonetoku, Ghirlanda, lag–luminosity, variability–luminosity, and Ep–α relations with sample selection, group fits, intrinsic scatter, and CSV export.
• Statistics — Mollweide sky distribution, redshift histograms, band coverage, and host-galaxy M*/SFR populations.

Filter management — 81 UV→far-IR filter definitions with transmission curves (SVO FPS), effective wavelengths, and Vega→AB conversions, feeding the extinction correction, synthetic photometry, and SED fits:

• Figure digitizer — extract points from published figures: axis calibration (linear/log), manual and color-based auto picking, CSV export or direct database write. An original reimplementation inspired by the MIT-licensed graph-digitizer, offered as a self-hosted alternative to WebPlotDigitizer (AGPL).
• GCN circular reader — browse the local GCN archive (45,000+ circulars) with per-source info cards, Δt calculator, and a photometry-entry form that writes directly into the database.
• Light-curve template library × K-correction — versioned templates with QC surfaces and error budgets, a template builder wizard, and K-corrected prediction curves overlaid on the multi-source comparison page (optional ChromaShift engine).
• Multi-source comparison — overlay any number of events with per-band selection and server-side CSV/JSON export.
• Pipeline ingest API — POST /api/ingest/photometry (Bearer token) for direct wiring to the STDpipe/STDweb photometry pipeline; a full REST API covers all entities (see docs/TECHNICAL.md).

The catalog lives in the separate AJST-Data repository (CC BY 4.0) and is optional — AJST runs on an empty database with your own data.

Data were collected from published papers and GCN circulars and inherit part of the Dainotti et al. (2024, MNRAS 533, 4023) sample; every subset documents its provenance and parsing scripts.

The catalog has not yet been audited entry by entry. A systematic per-source review is underway, supported by a claim-based queue (audit/state.tsv + tools/audit_queue.py) that lets multiple human or LLM-assisted reviewers verify sources in parallel without collisions. Until that review completes, do not use the data directly for serious scientific research. Audit contributions are welcome — see AJST-Data's CONTRIBUTING.md.

Requires Python ≥ 3.10 and PostgreSQL ≥ 14.

git clone https://github.com/Cooper-J2000/AJST.git
cd AJST

python3 -m venv .venv && source .venv/bin/activate
pip install -r requirements.txt
# optional: pip install "VegasAfterglow[mcmc]" corner matplotlib   # afterglow fitting
# optional: pip install pcigale                                    # host SED fitting (pcigale engine)

createdb ajst_catalog
export AJST_CATALOG_PASSWORD='choose-a-strong-password'   # REQUIRED

# optional: import the catalog
git clone https://github.com/Cooper-J2000/AJST-Data.git catadata
cd backend && python3 etl.py && cd ..

./backend/start.sh        # listens on 127.0.0.1:27101 (loopback only)

The layered test suite (847 unit, API-contract, and acceptance cases; 843 passed

• 4 skipped as of 2026-10-03) runs with python3 -m pytest tests/.

AJST builds on open tools and catalogs: dustmaps (CSFD map, Liu et al. 2023) and dust_extinction (Pei 1992); VegasAfterglow; CIGALE/pcigale (Boquien et al. 2019, A&A 622, A103); prospector (Johnson et al. 2021, ApJS 254, 22) with python-FSPS; the SVO Filter Profile Service; Aladin Lite v3 (CDS); the Transient Name Server (spectral-line lists); Bootstrap, Chart.js, and bootstrap-icons (bundled, MIT); and the GCN archive (NASA), HEASARC/MAST catalogs (BATSE, Fermi GBM/LAT, Swift BAT/XRT/UVOT, Konus-Wind, AGILE-MCAL), the GRBSN webtool, and literature samples (Dainotti et al. 2024; Wang et al. 2022; Minaev & Pozanenko 2020; Guidorzi et al. 2025; Liang et al. 2023; Chandra & Frail 2012). Per-directory READMEs in AJST-Data list the exact provenance and citation requirements for each data subset.

Issues and pull requests are welcome (see docs/COMMIT-CONVENTION.md); for data auditing, start from AJST-Data's CONTRIBUTING.md. Code is released under the MIT License; the data under CC BY 4.0, with external/ subsets subject to their original sources' terms and citation requirements.